Weizmann Institute of Science


Advanced Topics In Computer Vision
And Deep Learning

Spring 2021

[ Home | Schedule | Prerequisites | Useful Links ]

The full topics spreadsheet is available HERE

The zoom recordings are only available with Weizmann credentials

Date Speakers Papers Slides Zoom Recording
Introduction Meeting
21/3/2021 Michal Irani
Niv Haim
Shai Bagon
Course Intro (and DOs and DON'Ts when giving a talk)
Talk Preparation Timeline and Guidelines
Zoom intro
Attention/Transformers in Vision
4/4/2021 Shir Amir
Shiran Zada
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
On the Relationship between Self-Attention and Convolutional Layers
Taming Transformers for High-Resolution Image Synthesis
Stand-Alone Self-Attention in Vision Models
Is Space-Time Attention All You Need for Video Understanding?
Training data-efficient image transformers & distillation through attention
Slides Zoom
Advances in Reinforcement Learning
11/4/2021 Hodaya Koslowski
Yuval Belfer
Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
Mastering Atari with Discrete World Models
Combining Deep Reinforcement Learning and Search for Imperfect-Information Games
Slides Zoom
StyleGAN and Applications
18/4/2021 Or bar-Shira
Oz Frank
Progressive Growing of GANs for Improved Quality, Stability, and Variation
A Style-Based Generator Architecture for Generative Adversarial Networks
Stylegan2: Analyzing and Improving the Image Quality of StyleGAN
Stylegan-Ada: Training Generative Adversarial Networks with Limited Data
Interpreting the Latent Space of GANs for Semantic Face Editing
Transforming and Projecting Images into Class-conditional Generative Networks
Time-Travel Rephotography
DeepLandscape: Adversarial Modeling of Landscape Videos
Slides Zoom
Image Synthesis without GANs
25/4/2021 Itai Antebi
Rafail Fridman
StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows
SRFlow: Learning the Super-Resolution Space with Normalizing Flow
NVAE: A Deep Hierarchical Variational Autoencoder
Generating Diverse High-Fidelity Images with VQ-VAE-2
Glow: Generative Flow with Invertible 1x1 Convolutions
Slides Zoom
Deep Learning meets Natural Sciences
02/05/2021 Asaf Petruschka
Dror Bar
AlphaFold 2 & Equivariance
BERTology Meets Biology: Interpreting Attention in Protein Language Models
Slides Zoom
Efficient Attention Mechanisms
09/05/2021 Dana Joffe
Roy Abel
Reformer: The Efficient Transformer
Linformer: Self-Attention with Linear Complexity
Rethinking Attention with Performers
Efficient Content-Based Sparse Attention with Routing Transformers
Perceiver: General Perception with Iterative Attention
Generating Long Sequences with Sparse Transformers
Slides Zoom
Advances in Object Detection / Segmentation
23/05/2021 Ophir Sarusi
Raz Yerushalmi
DETR: End-to-End Object Detection with Transformers
CornerNet: Detecting Objects as Paired Keypoints
Pixel Consensus Voting for Panoptic Segmentation
YOLACT: Real-time Instance Segmentation
HyperSeg: Patch-wise Hypernetwork for Real-time Semantic Segmentation
YolactEdge: Real-time Instance Segmentation on the Edge
Slides Zoom
Normalization Techniques and its importance in Generative Models
30/05/2021 Sveta Paster
Ore Shtalrid
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Instance Normalization: The Missing Ingredient for Fast Stylization
Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization
SPADE: Semantic Image Synthesis with Spatially-Adaptive Normalization
How Does Batch Normalization Help Optimization?
High-Performance Large-Scale Image Recognition Without Normalization
Slides Zoom
Neural Implicit Representations
06/06/2021 Eyal Naor
Dolev Ofri
Occupancy Networks: Learning 3D Reconstruction in Function Space
DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation
Implicit Neural Representations with Periodic Activation Functions
Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
NeX: Real-time View Synthesis with Neural Basis Expansion
Slides Zoom
Meta Learning
13/06/2021 Lior Magram
Bar Karov
Fast Adaptation to Super-Resolution Networks via Meta-Learning
Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
Learned Initializations for Optimizing Coordinate-Based Neural Representations
On First-Order Meta-Learning Algorithms
Slides Zoom
Deep Denoising
20/06/2021 Alon Mamistvalov
Hila Naaman
Denoising and Regularization via Exploiting the Structural Bias of Convolutional Generators
Robust And Interpretable Blind Image Denoising Via Bias-Free Convolutional Neural Networks
Slides Zoom
Depth Estimation from Single Image
27/06/2021 Narek Tumanyan
Navve Wasserman
Unsupervised Monocular Depth Estimation with Left-Right Consistency
Unsupervised Learning of Depth and Ego-Motion from Video
Consistent Video Depth Estimation
Robust Consistent Video Depth Estimation
Slides Zoom
Out of Distribution Detection
04/07/2021 Myriam Schmidt A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks
Slides Zoom